Re: Horrible term-mode problems on upgrade to 21.4
Nix <$}xinix{$(a)esperi.demon.co.uk> writes:
All very well, except that in term mode, CR/LF's come in and are
interpreted by `term-emulate-terminal' in a DOSsish fashion, with CR and
LF serving their ancient teletype purposes; as this is how TTYs work,
this is entirely correct.
What is *not* correct is XEmacs picking up on the CR/LFs and translating
them to LFs!
Thanks for the analysis. I've successfully repeated the bug in my
Mule-enabled XEmacs, and I see the same problem. A quick fix that
worked for me is to evaluate this from another (whichever) buffer:
(with-current-buffer (get-buffer "*terminal*")
(set-buffer-process-coding-system 'binary 'binary))
This sets the "process coding system" in the "*Terminal*" buffer to
`binary', which means "no conversion whatsoever, thankyouverymuch".
I'd fix term-mode, if it wasn't that setting the
`buffer-file-coding-system' to `no-conversion'
That's because `no-conversion' actually means "no charset
conversion". Line endings still get converted. (Yes, it's
confusing. It will probably be changed in one of the future
releases.)